Topic portrait system and topic portrait method based on Zhihu

Abstract

The invention discloses a topic portrait system and topic portrait method based on Zhihu data. The system includes a data preprocessing module for extracting, cleaning and preprocessing data from a website, a topic portrait module for accurately portraiting a topic, and A user graphical interface module for visually presenting the results of the topic portrait module and downloading reports; the topic portrait method includes the following steps: (1) extracting, cleaning, and preprocessing data from the website, specifically including topic data crawling, data cleaning and Preprocessing; (2) Precise portrait of the topic, specifically including data statistical analysis, user portrait analysis, network data analysis, text data analysis, labeling the specified features of the topic and comparative analysis with similar topics; (3) Presentation of the user image interface, specifically Including visual presentation of analysis results and downloading of analysis reports; the present invention presents portrait results through an interactive and visualized user image interface, which broadens the mining and application of Zhihu data.

Description

technical field [0001] The present invention relates to an artificial intelligence system and a data mining method, in particular to a topic portrait system based on Zhihu, and also to a topic portrait method based on Zhihu. Background technique [0002] Zhihu is an emerging online question-and-answer community in recent years, with the concept of sharing each other's professional knowledge and experience, and maintaining a rigorous and rational community atmosphere. As of September 2017, Zhihu has more than 100 million individual registered users and 18 billion monthly views. Zhihu integrates social elements on the basis of the question-and-answer community, redefines the relationship between people and information, and establishes a new content creation and dissemination mechanism. Its high-quality community content has gradually become an important way for Internet users to acquire knowledge. [0003] Zhihu topic is a kind of social tagging (Social Tagging). Users create...

AI Technical Summary

Problems solved by technology

[0004] Judging from the existing research results, although the network Q&A community information has been used in user behavior research, information quality research, knowledge dissemination research and decision support research, most of the research objects are still in the first generation of keyword search Q&A. Among them, the research of information science and technology focuses on algorithm optimization, and the research of social science is mainly qualitative and experience summarization. There is no content structure and dissemination mechanism for Zhihu, and systematic and universal data mining around topics. method

At the same time, in the research of the second-generation social Q&A community, topic research mainly focuses on topic identification, focusing on natural language processing, including technologies such as network information capture, natural language segmentation, and keyword extraction. There are also the following limitations : 1. Research mainly focuses on methods such as topic and keyword extraction, sentiment analysis, etc., providing information with a single dimension and not being associated with specific application scenarios

2. The application of natural language processing technology is still immature, and many shortcomings such as complex Chinese semantics and lack of training corpus make the actual performance of natural language processing technology poor

3. Most of the text data in the online question-and-answer community is a collection of short texts, which has the characteristics of fragmentation, colloquial expression, and sparse data sets, which brings new challenges to natural language processing technology

Images